From: | Grzegorz Jaśkiewicz <gryzman(at)gmail(dot)com> |
---|---|
To: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> |
Cc: | pgsql-hackers(at)postgresql(dot)org |
Subject: | Re: small fix to possible null pointer dereference in byteaout() varlena.c |
Date: | 2010-09-28 15:15:13 |
Message-ID: | AANLkTi=u9vVNgeGeXaGpkA-FjDQjBu_2wp+GNWHD0ihK@mail.gmail.com |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-hackers |
2010/9/28 Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>:
> =?UTF-8?Q?Grzegorz_Ja=C5=9Bkiewicz?= <gryzman(at)gmail(dot)com> writes:
>> It would crash if input is of unrecognized format. Probably than
>> there's going to be more problems to be concerned with, but just in
>> case, don't crash in
>
> I'm not sure why you think this is a good change, but it would break
> things: in particular, the code would fail to null-terminate the string
> in the hex-output case. Also, the case that you seem to be trying to
> defend against can't happen because elog(ERROR) doesn't return.
>
...
rp = result = NULL; /* keep compiler quiet */
}
*rp = '\0';
....
this strikes me as a clear case of possible null pointer dereference,
wouldn't you agree ?
I know the case is very corner-ish, but still valid imo.
--
GJ
From | Date | Subject | |
---|---|---|---|
Next Message | Robert Haas | 2010-09-28 15:23:32 | Re: security label support, revised |
Previous Message | Tom Lane | 2010-09-28 15:14:43 | Re: security label support, revised |